Research On Edge Computing And Collaborative Offloading Mechanism Based On SDN In Vehicular Networks

With the development of the Internet of Things and intelligent transportation systems,the amount of tasks generated in real-time in the traffic environment has proliferated.At the same time,the diversity and complexity of tasks have become more and more demanding on the computing resources and storage resources of in-vehicle computing devices.The introduction of cloud computing accelerates the processing of vehicle intensive tasks to meet user needs.However,the cloud-based approach also has the following problems: First,the core cloud server is usually deployed at the back end of the backbone network,and remote cloud offloading will result in a longer.The transmission backhaul increases the communication overhead and transmission delay.Secondly,the load of the cellular base station is too large in the dynamic vehicle-intensive network,resulting in low data transmission efficiency.Third,the real-time vehicle computing task is unevenly distributed during the processing,making the service Congestion and system overhead increase.In response to these questions,this paper has carried out the following research:(1)An architecture for vehicle task offloading in a dynamic traffic scenario is proposed,namely a software defined vehicle edge computing architecture.Through the integration of software to define the advantages of global control in the network and the advantages of fast execution under multi-access edge computing,it provides a flexible and controllable system framework for vehicle data offloading and computational load sharing.(2)A vehicle data task shunting algorithm based on the maximum connection aging model is proposed.The longest communication path planning and data task shunting are realized by software-defined collaborative control of the vehicle networking controller.Firstly,this paper designs the aging model of communication connection between vehicle and vehicle,vehicle and roadside unit,and the aging model of communication path between multiple vehicles.Secondly,the software-defined vehicle networking controller is modified by improving Bellman-ford algorithm.The vehicle information table performs a search comparison to obtain the longest time-sensitive communication path between the source vehicle and the destination vehicle;finally,by globally recognizing the traffic scene,the road future information is predicted and the command can be quickly executed.The proposed model and algorithm can effectively share traffic from the base station,reduce communication and control overhead,and improve the efficiency of data delivery.(3)An offloading algorithm for vehicle calculation type tasks with joint delay and energy consumption is proposed.In this paper,the delay and energy consumption model of vehicle computing type task to edge server is designed.At the same time,the cost model of task migration in the server is designed for mobile vehicles.The minimum overhead is the target,and the calculation and communication in the joint task processing.Delay and energy consumption for task offloading;Optimize Hopcroft-karp algorithm to optimize vehicle and edge server matching to obtain maximum matching under minimum cost,ensure maximum utilization of server resources;collect global information to the edge through controller.The server issues commands to improve the efficiency of task assignment and virtual machine migration.This paper builds a vehicle network information physics system simulation verification platform based on SUMO+NS3 and OpenDaylight+Mininet architecture,and effectively validates and analyzes the research of vehicle data type task shunt and computation type task load.In the simulation of the data task offloading scheme,the two schemes are compared with the other two schemes,and the task's splitting rate and transmission delay are evaluated.In the simulation of the computing task load,the load sharing rate,delay and energy consumption of the migration model are compared.The simulation results show that the SDN collaborative optimization algorithm proposed in this paper can effectively improve the load sharing rate and reduce the load sharing delay and energy consumption.